Deconvolution as a feasible alternative to the root-sum-squares rule for non-Gaussian transition duration analysis

A 100-ps Debye line filter is analyzed for transition duration using both deconvolution and root-sum-squares methods. The latter method shows errors of ±12 percent for the cases presented, while deconvolution agrees to within 0.3 percent of the theoretical filter model. In addition, the deconvoluted Debye filter waveform matches the theoretical step response to within less than 1 percent at all points in a 500-ps time epoch.